Logical Foundations for Cognitive Agents